Human-Object Interaction Generation on BEHAVE (test)
0FIDGround Truth
Evaluation Results
| Method | Links | |||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ground TruthBatch size=2562026.03 | 0 | 55.9 | 91 | 92.6 | 1.483 | 6.94 | — | — | 18.1 | — | 9.4 | 20.4 | 100 | 100 | 100 | — | — | — | — | |
| Real motion2025.03 | 0.001 | 28.7 | 44.3 | 54.4 | 1.055 | 1.346 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ChainHOIPost-processing (AIC)=true2025.03 | 0.093 | 44.4 | 62.3 | 72.2 | 0.964 | 1.337 | 0.072 | 0.099 | 5.8 | 0.078 | — | — | — | — | — | — | — | — | — | |
| ChainHOI2025.03 | 0.095 | 43.5 | 62.1 | 71.7 | 0.967 | 1.337 | 0.091 | 0.081 | 6.3 | 0.096 | — | — | — | — | — | — | — | — | — | |
| CHOISvariant=Star2025.03 | 0.157 | 30.1 | 48.8 | 60.6 | 1.09 | 1.265 | 0.187 | 0.086 | 11.8 | 0.202 | — | — | — | — | — | — | — | — | — | |
| InerDiff2025.03 | 0.17 | 31 | 48 | 59.9 | 1.045 | 1.325 | 0.191 | 0.078 | 6.9 | 0.206 | — | — | — | — | — | — | — | — | — | |
| MDMFinetuning=true2025.03 | 0.246 | 22.3 | 37.8 | 48.8 | 1.118 | 1.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MDMvariant=Star2025.03 | 0.257 | 22 | 35.5 | 45.1 | 1.071 | 1.307 | 0.458 | 0.095 | 9.8 | 0.481 | — | — | — | — | — | — | — | — | — | |
| PriorMDMvariant=Star2025.03 | 0.328 | 24.3 | 32.9 | 38.5 | 1.203 | 1.142 | 0.215 | 0.116 | 6.6 | 0.232 | — | — | — | — | — | — | — | — | — | |
| HOI-DiffPost-processing (AIC)=true2025.03 | 0.437 | 31.2 | 46.7 | 56.3 | 1.107 | 1.235 | 0.101 | 0.081 | 9.8 | 0.117 | — | — | — | — | — | — | — | — | — | |
| HOI-Diff2025.03 | 0.457 | 29.5 | 44.1 | 53.9 | 1.119 | 1.204 | 0.148 | 0.102 | 12.5 | 0.214 | — | — | — | — | — | — | — | — | — | |
| LIGHTGuidance=true, Batch size=2562026.03 | 0.481 | 27.7 | 47.5 | 56.1 | 3.277 | 6.961 | — | — | 17.1 | — | 14 | 17 | 73.1 | 72 | 69.7 | — | — | — | — | |
| LIGHTGuidance=false, Batch size=2562026.03 | 0.5 | 22.5 | 43 | 51.8 | 3.379 | 6.92 | — | — | 16.3 | — | 14.3 | 15.8 | 72.2 | 69.3 | 67.6 | — | — | — | — | |
| InterDiffBatch size=2562026.03 | 0.63 | 25.8 | 44.3 | 53.1 | 3.219 | 6.862 | — | — | 20.8 | — | 14.3 | 17.3 | 72.3 | 68.2 | 67.6 | — | — | — | — | |
| HOI-DiffBatch size=2562026.03 | 0.689 | 41.3 | 62.4 | 74 | 3.029 | 7.62 | — | — | 7.2 | — | 10.3 | 8.4 | 72.2 | 44.7 | 50.1 | — | — | — | — | |
| CHOISBatch size=2562026.03 | 0.705 | 21.9 | 39.6 | 49.2 | 3.423 | 6.863 | — | — | 19.1 | — | 12.8 | 17 | 71.7 | 66.2 | 65 | — | — | — | — | |
| IMAGIN-4Dimage-conditioned=true, reference=MeshImg2026.06 | 0.76 | — | — | — | — | — | — | — | — | — | — | — | 53.8 | 38.1 | 39.7 | 12.4 | 12.23 | 11.18 | 4.93 | |
| Text2HOIBatch size=2562026.03 | 1.291 | 13.5 | 24.2 | 32.2 | 3.815 | 6.544 | — | — | 17.7 | — | 17.9 | 12.8 | 70.3 | 57.9 | 58.7 | — | — | — | — | |
| CHOISimage-conditioned=false2026.06 | 1.39 | — | — | — | — | — | — | — | — | — | — | — | 52.6 | 35.1 | 37.5 | — | — | — | 3.05 | |
| ViHOI*image-conditioned=true, reference=MeshImg, adaptation=single-image2026.06 | 1.44 | — | — | — | — | — | — | — | — | — | — | — | 51.8 | 34.8 | 37.8 | 22.97 | 20.34 | 14.55 | 4.49 | |
| CHOIS+Imgimage-conditioned=true, reference=MeshImg2026.06 | 1.57 | — | — | — | — | — | — | — | — | — | — | — | 52.1 | 36.1 | 38.6 | 22.02 | 18.81 | 13.71 | 4.07 |